USAGE OF DYNAMIC LOAD BALANCING FOR DISTRIBUTED SYSTEM IN CLOUD COMPUTING

USAGE OF DYNAMIC LOAD BALANCING FOR DISTRIBUTED SYSTEM IN CLOUD COMPUTING Reeta Mishra 1 Assistant Professor, K.J.Institute of Engineering & Technology,Savli,Vadodara,Gujarat (India) ABSTRACT Cloud computing involves virtualization, distributed computing,networking, software and web services. A cloud consists of several elements suchas clients, datacenter and distributed servers. It includes fault tolerance, high availability,scalability, flexibility, reduced overhead for users, reduced cost of ownership, ondemand services etc. Two of the main obstacles in the usage of cloud computing are Cloud Security and Performance stability.performance stability can be improved byload balancing.load balancing ensures that all the processor in the system orevery node in the network does approximately the equal amount of work at any instantof time.this technique can be sender initiated, receiver initiated or symmetric type(combination of sender initiated and receiver initiated types). Keywords: Load Balancing, Cloud Computing, Cloud Security, Performance Stability, Distributed System. I. INTRODUCTION Cloud computing is an on demand service in which shared resources, information, software and other devices are provided according to the clients requirement at specific time. It s a term which is generally used in case of Internet. The whole Internet can be viewed as a cloud. Capital and operational costs can be cut using cloud computing. Cloud computing is a model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, data storage, software applications and other computing services) that can be rapidly provisioned and released with minimal management effort or service provider interaction. Figure 1: A Cloud is Used in Network Diagrams to Depict the Internet [1]. There are three services of cloud shown as below- 625 P a g e

1) Infrastructure as a service (IaaS) 2) Software as a service (SaaS) 3) Platform as a service (PaaS) Figure 2: Cloud services II. DISTRIBUTED FILE SYSTEM Distributed file system for cloud is a file system that allows many clients to have access to the same data/file providing important operations (create, delete, modify, read and write). Each file may be partitioned into several parts called chunks. Each chunk is stored in remote machines. Typically, data is stored in files in a hierarchical tree where the nodes represent the directories. Hence, it facilitates the parallel execution of applications. There are several ways to share files in a distributed architecture. Meanwhile, the security and performance of the system must be ensured. Confidentiality, availability and integrity are the main keys for a secure system. Nowadays, users can share resources from any computer/device, anywhere and everywhere through internet thanks to cloud computing which is typically characterized by the scalable and elastic resources such as physical servers, applications and any services that are virtualized and allocated dynamically. Thus, synchronization is required to make sure that all devices are update. Distributed file systems enable also many big, medium and small enterprises to store and access their remote data exactly as they do locally, facilitating the use of variable resources. III. LOAD BALANCING Load balancing in cloud computing systems is really a challenge now. It means distributing the amount of work to do between different servers in order to get more work done in the same amount of time and serve clients faster. Always a distributed solution is required. Because it is not always practically feasible or cost efficient to maintain one or more idle services just as to fulfill the required demands. Jobs can t be assigned to appropriate servers and clients individually for efficient load balancing as cloud is a very complex structure and components are present throughout a wide spread area. Here some uncertainty is attached while jobs are assigned. 626 P a g e

IV. CHALLENGES IN CLOUD COMPUTING LOAD BALANCING Finding a solution for problems in load balancing is never an easy process there will bemany challenges to be faced while developing a solution. Here in this section we will discusssome of the common challenges that might be faced while developing a solution for a problemof load balancing in cloud computing. a) Distribution of Cloud Nodes: There are many algorithms being proposed for loadbalancing in cloud computing. Among them some algorithms might produce efficient results withsmall networks or a network with closely located nodes. Such algorithms are not suitable forlarge networks because those algorithms cannot produce the same efficient results whenapplied to larger networks. There are many reasons that affect the efficiency in larger networkslike speed of the network, distance between the clients and server nodes and also the distancebetween all the nodes in the network [2]. So while developing a load balancing algorithm oneshould try for better results in spatially distributed nodes balancing the load effectively reducing network delays. b) Migration Time: In cloud computing the service-on-demand method will be followedwhich means when there is a demand for a resource the service will be provided to the requiredclient. So while serving the client on his demands sometimes we need to migrate resourcesfrom long distances due to unavailability in near locations. In such cases the time of migration ofthe resources from far locations will be more which will affect the performance of the system.while developing an algorithm one should note that resource migration time is an importantfactor that greatly affects the performance of the system. c) Performance of the System: It doesn t mean that if the complexity of an algorithm ishigh then the performance of the system will be high. Any time load balancing algorithm must besimple to implement and easy to operate. If the complexity of algorithm is high then theimplementation cost will also be more and even after implementing the system performance willbe decreased due to more delays in the functionality of the algorithm. d) Failure of controller: Definitely centralized load balancing algorithms (Having onecontroller) can provide efficient results while balancing the load than the distributed algorithms.but in centralized load balancing algorithms when the controller fails the whole system will behalted, in such cases there will be a huge loss for both client and service provider. So, the loadbalancing algorithms must be designed in a decentralized and distributed fashion so that whena node acting as a controller fails the system will not halt [5]. In such cases the control will begiven to other nodes and they will act as controllers of the system. e) Energy Management: A load balancing algorithm should be designed in a way such thatthe operational cost and the energy consumption of the algorithm must be low. Increase in theenergy consumption is one of the main problems that cloud computing is facing today. Eventhough by using energy efficient hardware architectures which slows down the processor speedand turn off machines that are not under use the energy management is becoming difficult. So to achieve better results in energy management a load balancing algorithm should be designedby following Energy Aware Job Scheduling methodology [6]. f) Security: Security is one of the problems that cloud computing has in its top mostpriority. The cloud is always vulnerable in one or the other way to security attacks like DDOSattacks etc. While balancing the load there are many operations that take place like VMmigrationetc at that time there is a high probability 627 P a g e

of security attacks. So an efficient loadbalancing algorithm must be strong enough to reduce the security attacks but should not bevulnerable. V. DYNAMIC LOAD BALANCING ALGORITHM In the distributed one, the dynamic load balancing algorithmis executed by all nodes present in the system and the task of load balancing is sharedamong them. The interaction among nodes to achieve load balancing can take two forms: cooperative and non-cooperative [4]. In the first one, the nodes work side-by-side to achieve a common objective, for example, to improve the overall response time, etc. In the second form, each node works independently toward a goal local to it, for example, to improve the response time of a local task. Dynamic load balancing algorithms of distributed nature, usually generate more messages. A benefit, of this is that even if one or more nodes in the system fail, it will not cause the total load balancing process to halt, it instead would effect the system performance to some extent. Distributed dynamic load balancingcan introduce immense stress on a system in which each node needs to interchangestatus information with every other node in the system. It is more advantageous when most of the nodes act individually with very few interactions with others. 5.1 Policies or Strategies in dynamic load balancing There are 4 policies [4]: Transfer Policy: The part of the dynamic load balancing algorithm which selects a job for transferring from a local node to a remote node is referred to as Transferpolicy or Transfer strategy. Selection Policy: It specifies the processors involved in the load exchange (processormatching) Location Policy: The part of the load balancing algorithm which selects a destination node for a transferred task is referred to as location policy or Location strategy. Information Policy: The part of the dynamic load balancing algorithm responsible for collecting information about the nodes in the system is referred to as Information policy or Information strategy. Figure3 : Interaction among components of a dynamic load balancing algorithm 628 P a g e

VI. HONEYBEE FORAGING ALGORITHM In case of load balancing, as the webservers demand increases or decreases, the services are assigned dynamically to regulate the changing demands of the user. The servers are grouped under virtual servers (VS), each VS having its own virtual service queues. Each server processing a request from its queue calculates a profit or reward, which is analogous to the quality that the bees show in their waggle dance. One measure of this reward can be the amount of time that the CPU spends on the processing of a request. The dance floor in case of honey bees is analogous to an advert board here. This board is also used to advertise the profit of the entire colony. Each of the servers takes the role of either a forager or a scout. The server after processing a request can post their profit on the advert boards with a probability of pr. Aserver can choose a queue of a VS by a probability of px showing forage/explore behavior, or it can check for advertisements (see dance) and serve it, thus showing scout behavior. A server serving a request, calculates its profit and compare it with the colony profit and then sets its px. If this profit was high, then the server stays at the current virtual server; posting an advertisement for it by probability pr. If it was low, then the server returns to the forage or scout behavior. Figure 4: Server Allocations by Foraging in Honey bee technique VII. BIASED RANDOM SAMPLING Here a virtual graph is constructed, with the connectivity of each node (a server is treated as a node) representing the load on the server. Each server is symbolized as a node in the graph, with each indegree directed to the free resources of the server. Regarding job execution and completion, Whenever a node does or executes a job, it deletes an incoming edge, which indicates reduction in the availability of free resource. After completion of a job, the node creates an incoming edge, which indicates an increase in the availability of free resource. 629 P a g e

VIII. ACTIVE CLUSTERING Active Clustering works on the principle of grouping similar nodes together and working on these groups. The process involved is: A node initiates the process and selects another node called the matchmaker node from its neighbors satisfying the criteria that it should be of a different type than the former one. The so called matchmaker node then forms a connection between a neighbor of it which is of the same type as the initial node. The matchmaker node then detaches the connection between itself and the initial node. IX. PROPOSEDWORK The distributed network may follow different topologies. The tasks are distributed over the whole network. One topological network connects with the other through a gateway. One of the physical topologies forming a cloud. in case of clouds is an optimal division of loads among a number of master computers, slave computers and their communication links. Our objective is to obtain a minimal partition of the processing load of a cloud connected via different communication links such that the entire load can be distributed and processed in the shortest possible amount of time. The concept of load balancing in Wireless sensor networks (WSN) proposedin [9] can also be applied to clouds as WSN is analogous to a cloud having no. ofmaster computers (Servers) and no. of slave computers (Clients).The slave computers are assumed to have a certain measurement capacity. We assume that computation will be done by the master computers, once all the measured data isgathered from corresponding slave computers. Only the measurement and communicationtimes of the slave computers are considered and the computation time of the slave computersis neglected. Here we consider both heterogeneous and homogeneous clouds. That isthe cloud elements may possess different measurement capacities, and communication linkspeeds or the same measurement capacities, and communication link speeds. One slavecomputer may be connected to one or more master computers at a certain instant of time.in case of clouds, an arbitrarily divisible load without having any previousrelations is divided and first distributed among the various master computers (for simplicity here the load is divided equally between the master computers) and the each master computer distributes the 630 P a g e

load among the corresponding slave computers so that the entire load can be processed in shortest possible amount of time. X. CONCLUSION In this paper we discussed the significance of load balancing in cloud computing andalso we discussed various challenges that occur while balancing load in a cloud computingnetwork. Performance of the algorithms which were being implemented for balancing the load in cloud computing was not up to the requirements of cloud. There are different areas involved in achieving the load balancing of a cloud. The problem that we are facing with the existing load balancing algorithms is they are not able to perform well in all the required areas of load balancing. REFERENCES [1] Anthony T.Velte, Toby J.Velte, Robert Elsenpeter, Cloud Computing A Practical Approach,TATAMcGRAW-HILL Edition 2010. [2] Ali M. Alakeel, A Guide to Dynamic Load Balancing in Distributed Computer Systems, IJCSNS International Journal of Computer Science and Network Security, VOL.10 No.6, June 2010. [3] wikipedia.org, http://en.wikipedia.org/wiki/load_balancing_(computing). [4] V Krishna Reddy, Srikanth Reddy. A Survey of Various Task Scheduling Algorithms in CloudComputing. i-manager s Journal on Computer Science (JCOM). 2013; 1(1). [5] http://www-03.ibm.com/press/us/en/pressrelease/22613.wss [6] Shanti swaroopmoharana, Rajadeepan d Ramesh, Digamberpowar. Analysis Of Load Balancers In Cloud Computing. International Journal of Computer Science and Engineering (IJCSE). ISSN 2278-9960. 2013; 2(2). [7] Martin Randles, EnasOdat, David Lamb, Osama Abu- Rahmeh and A. Taleb-Bendiab, A Comparative Experiment in Distributed Load Balancing, 2009 Second InternationalConference on Developments in esystems Engineering. [8] S. Dhakal, B. S. Paskaleva, M. M. Hayat, E. Schamiloglu, C. T. Abdallah, Dynamical Discrete-Time Load Balancing in Distributed Systems in the presence of time delays, Decision and Control, 2003. Proceedings. 42nd IEEE Conference, Dec 2003, Volume:5, Pages:5128-5134. [9] MequanintMoges, Thomas G.Robertazzi, Wireless Sensor Networks: Scheduling for Measurement and Data Reporting, August 31, 2005 631 P a g e

PROPOSAL OF A MODIFIED INCENTIVE SCHEME FOR THE GROWTH OF AN ORGANIZATION Riju Thomas 1, Dr. Biju Augustine P. 2 1 Student, 2 Associate Prof, Government Engineering College Rajiv Gandhi Institute of Technology ABSTRACT It is a universal fact that everyone works for result. In an industry, if a definite amount is set as salary then employees will work only till they reach the target. But once an incentive scheme is proposed whereby an employee knows more the work done (after achieving the target) more will be the incentive, he/she will be more loyal to the industry which will result in higher production. In this paper point based incentive scheme is discussed and a better incentive scheme called Novartis scheme is implemented based on various situations. Keywords: Employees, Incentive, Novartis, Production, Point based, Target. I. INTRODUCTION It is a known fact that everyone has a purpose behind every act. The same fact implies for employees working in various companies or industries. In this case, the purpose is salary. Now that in most cases, employee is already told the definite salary that he/she will earn after a month or year, it is known that he/she will become casual to the work as salary remains constant. [1] It is a known fact that motivational concept are central to employees. So came the concept of incentive, whereby apart from salary, employees are offered incentives on the basis of amount of work done after achieving a definite target set by the organization. [2] Incentive is a means by which an organization motivates and increases the performance of its workers. [3] Motivation and performance are greater when chronic predispositions, task incentives, and means of goal attainment all share the same regulatory focus than when they do not. [4] The two characteristics that influences the work motivation are job characteristics and work content. [5] Every employee wants a caring organization; benefits can be source of competitive advantage. So incentives can be termed as hope that an employee has, to be rewarded if he/she works efficiently and crosses the determined target. [6] Motivation must come from within the individual (Zacarelli, 1985; Simons & Enz, 1995; Nicholson, 2003). Incentive not only increases the productivity but also enhances commitment in the employees. [7] Incentives are also being offered to study whether the individual religious performance increases. [8] Incentive is the offer of a reward before the action is taken so that the desired behaviour can be induced. Incentive is a best method for the employer to show that the input of the employees is valued, [9] as it is always efficient for the firm to retain the worker in all states of the world. When broadly classified, incentives may be monetary or non monetary. In monetary incentive scheme, employee is given monetary benefits for his/her dedicated work like bonus. [10] Monetary incentive may change how tasks are being perceived but if the incentives are not large enough; this change in perception can lead to undesired effects on behaviour. [11] When individuals are assigned tasks for which they do not have the 632 P a g e

skills, they wont increase the effort under monetary incentives as they believe that effort wont lead to increase in performance and consequent rewards. In non monetary incentive scheme, as the term suggests, employee is given benefits but not in monetary terms which [12] includes praises and promotions (Beardwell, Holden, & Claydon, 2004; Hoque, 2003). There are various reasons why an industry require incentive scheme which is explained below: Increase objective achievement rate If the employees are offered more money for more well defined i.e. qualitative and quantitative work, employer can set goals that are realistic because there are full chances that those objectives will be achieved as employees are rewarded for their hard work. Less number of absentees If the employee is set a target and if he/she knows that after crossing the target, monetary benefits may be availed, so he/she may try to be present in the industry for more number of days so as to cross the target. Mutual benefits Incentive scheme is a scheme where both the parties i.e. the employer and the employee are given the benefits. Employer is benefitted as target is reached and sometimes target is crossed as well. Employees are benefitted in terms of monetary incentives. Increase in production It may be clear from the above mentioned points that when monetary incentive is given to the employees on crossing the target, more and more employees will try and cross the target. Thus, in other terms Productivity will increase. Increased loyalty rate When employees know that the industry is rewarding them for their work, they too will be loyal to the company. If the employees of the industry are loyal, then and only then can an industry reach great heights. So, it may be stated that setting incentive schemes in an industry is not only beneficial for the employees but also for the employer or in other words the industry as a whole. There are various incentive schemes (inside monetary incentive schemes) that are used by various industries. One such scheme is point based incentive scheme which is explained below. II. POINT BASED INCENTIVE SCHEME Point based incentive scheme is a scheme, whereby an employee is rewarded points for his/her work once the target is crossed. Based on the points, incentive is decided. This point based incentive scheme is a crisp scheme. That is say for example an employee who produces 50 units more than target will get 20 points and the one who produces 51 units more than target will get 40 points, if the point recorded table states 30 to 50 units will yield 20 points and 51 above yield 40 points. That means there is a crisp line at 50 units beyond which point is 40 and beneath which point is 20. When these points are converted to monetary terms then say if Rs 50 per point is the ratio of incentive, then the person who produced 50 units will get Rs 1000 as incentive and the one who produced 51 units will get Rs 2000 as incentive. But if we consider real time scenario, the person who produced 51 units hasn t done that much of work for which he/she should be awarded incentive double the amount than that of the one who produced 50 units. 633 P a g e

When considering the case of industry, work is done in groups and it is the group which is credited or blamed for the work or the output of that stage and not any individual. [13] If the money is paid to the group then how the money is distributed among them can have a major impact on the type or amount of behavioural change that occurs (Christianson et al, 2006; Young and Conrad, 2007). In point based scheme also points that are awarded are for a group or team not an individual. It is a known fact that in most cases, the team that is formed will have both type of people, i.e. people who are more dedicated and people who are less dedicated. Also in a group there are people who take more leave and there are people who are regular to office. If in a group same points are rewarded for all the members (free riding problem), then in a way it is unjust for the ones who are dedicated and regular to office. Similarly this scheme will act as a demoralizer for them. Its result will be that the dedicated and regular employees will also become less dedicated and irregular as the monetary value (in terms of incentive) is the same for all the members in a group. So, a better incentive scheme is proposed after altering a few parameters whereby the incentive will vary from member to member even if they belong to a same group. III. NOVARTIS SCHEME The Novartis scheme is adapted from the Novartis group and some amendments are made in it so as to suit the organization in which the point based incentive scheme is used. The NOVARTIS norms have been replaced by the norms that are suitable for the organization, so that a comparison can be done between the proposed incentive formula and the point based incentive scheme. Novartis scheme includes a few factors namely: basic pay, incentive percentage, business performance multiplier and individual performance multiplier. Based on these factors, a formula was formulated which is called Novartis formula which is given below: Fig. 1: Novartis Formula The basic salary and incentive percentage is decided by the management after negotiations with the employees. Business performance multiplier depends upon how well the organization perform during the time period and based on that, a fixed scale will be given as a multiplier which ranges from 0 to 1.5 of the target amount. Individual performance depends upon factor attendance and the target that is being achieved by the employees in the respective section. Based on this, a scale will be provided to each employee and will range from 0 to 1.5 of the target amount. So this incentive calculation formula will help in motivating the employees to work as their individual performance is also valued and no free riding problem will arise as in the case with point based incentive system. With the help of this formula it is possible to calculate the percentage of incentive that should be given to the employees based on the basic salary. 634 P a g e

The proposed formula is designed to pay the employees based on performance measured against the target set by the organization, the individual performance and the performance of the organization as well. Fig. 2: Proposed Formula The incentive percentage to be given to the employees is set by the organization, that is if the company decides to give 15 % as incentive then the first part will have the value 0.15. The second part of the equation gives the value of the individual performance based on factor like attendance and the target achieved. According to Novartis the value should range between 0 and 1.5. Business performance multiplier ranges between the value 0 and 1.5. This depends on the business performance, that is if business performance is poor as compared to previous year then a value of 0.5 is given. When business performs normally or similar to previous year then a value of 1 is given and if the business performs well compared to previous year then a value of 1.5 is given. Novartis mentioned that the combined value of both the multiplier must not exceed 2 unless exceptional performance is measured and therefore the range was selected between 0 and 1.5. 3.1 Advantages and Disadvantages Following are the advantages of Novartis formula: Employees will be motivated to see the organization as their own. Each individual will be rewarded based on their own performance and the organization has the liberty to set the incentive percentage. The free riding problem is also avoided as everyone will have to work other than the point based incentive system in which all the employees under a group were awarded equal points based on their achievement. If the employees work less then they will be given only the basic pay but the one who is willing to work for the betterment of the organization will be receiving good amount as incentive due to his individual performance. The business performance multiplier is only one part of the equation so employees will be receiving money in spite the poor performance of the company unlike the profit sharing incentive scheme in which the employees received nothing when the company performance was poor. Once the mentioned target is achieved by the employees then no quantitative measurement is done so there won t be any affect on the quality of the product, and as the individual and business performance is a factor for incentive the workers will be motivated to work with heart. Following are the disadvantages of Novartis formula: The drawback lies when the monitoring system fails or the one who monitors the individual performance shows bias on the employees. The factors that are considered to give value to the multipliers also need to be calculated precisely else the better person will be more discouraged and the wrong person will be enjoying the benefit. 635 P a g e

3.2 Implementation Implementation of this formula is done using four situations namely: Situation I: Direct application of values into the formula. Situation II: Raise the salary of the employees. Situation III: Change the percentage set by the management. Situation IV: Alter the value of Business performance. 3.2.1 Situation I On applying values directly in the formula based on the performance of the industry, following is the comparitative result (when compared with the old scheme which is the point based scheme) that is obtained: Fig. 3: Implementation of Novartis Formula 3.2.2 Situation II In this method, a proposal is made to increase the basic salary of the employees. Then based on the increased salary, the percentage of incentive to be given to the employees is calculated. A comparison is made between the previous incentive scheme and the proposed incentive formula which is shown below: Fig. 4: Implementation After Increasing the Basic Salary. 636 P a g e

3.2.3 Situation III In this method we change the norm that is, a proposal is made to increase the percentage set by the management from 15% to 30%. After that a comparison is made between the previous incentive scheme and the proposed incentive formula which is shown below: Fig. 5: Implementation After Changing Percentage set by Management from 15% to 30% 3.2.4. Situation IV In this method we change the values of the norm Business Performance. Then a comparison is made between the previous incentive scheme and the proposed incentive formula which is shown below: IV. CONCLUSION Fig. 6: Implementation After Changing Business Performance Factor. Incentive is the reward that an employer gives to the employee on his/her dedicated work. But if the incentive is same for all the members in the group then dedicated and regular employees may become less dedicated and irregular as everybody is awarded same incentive. So, a better incentive scheme is implemented after a few amendments in the formula whereby incentive not only depends on the target but also on individual performance of the employee. So, different members or employees of the same group may have different incentives based on his/her dedication and regularity to the industry. 637 P a g e

REFERENCES [1] Rajeswari Devadass, Employees Motivation in Organizations: An integrative literature Review, International Conference on Sociality and Economics Development, IPEDR, Vol.10, 2011. [2] Falola Hezekiah Olubusayo, Ibidunni Ayodotun Stephen, Olokundun Maxwell, Incentives Packages and Employees Attitudes to Work: A Study Of Selected Government Parastatals In Ogun State, South-West, Nigeria, International Journal of Research in Business and Social Science IJRBS Vol.3 No.1, 2014 ISSN: 2147-4478. [3] James Shah, E. Tory Higgins, and Ronald S. Friedman, Performance Incentives and Means:How Regulatory Focus Influences Goal Attainment, Journal of Personality and Social Psychology. 1998, Vol. 74, No. 2, 285-293. [4] Bradley E. Wright, Public-Sector Work Motivation:A Review of the Current Literature and a Revised Conceptual Model, Journal of Public Administration Research and Theory. 2001:4:559-586. [5] Edwinah Amah, Christine. A. Nwuche, Nwakaego Chukuigwe, Effective Reward and Incentive Scheme for Effective Organizations, Research Journal of Finance and Accounting. ISSN 2222-1697 (Paper) ISSN 2222-2847 (Online). Vol.4, No.13, 2013. [6] Catherine R. Johnson, Employee Motivation:A Comparison Of Tipped And Non-Tipped Hourly Restaurant Employees, A thesis submitted for the degree of Master of Science in the Rosen College of Hospitality Management at the University of Central Florida Orlando, Florida. [7] Teresa García-Muñoz, Incentives in religious performance: a stochastic dominance approach, Judgment and Decision Making, Vol. 5, No. 2, April 2010, pp. 176 181. [8] Mark Reed, Erica von Essen, Alister Scott, Mark Everard, Diane Mitchell, Incentives Tools Literature Review. [9] Paul Oyer, Why Do Firms Use Incentives That Have No Incentive Effects?, The Journal of Finance, Vol. Lix, No. 4, August 2004.pp. 1619-1649. [10] Uri Gneezy, Stephan Meier, and Pedro Rey-Biel, When and Why Incentives (Don t) Work to Modify Behavior, Journal of Economic Perspectives, Volume 25, Number 4, Fall 2011, Pages 191 210. [11] Sarah E Bonner, Geoffrey B Sprinkle, The effects of monetary incentives on effort and task performance: theories, evidence, and a framework for research, Accounting, Organizations and Society 27 (2002) 303-345.www.elsevier.com/locate/aos. [12] Sindri Thor Hilmarsson, Pall Rikhardsson, The Evolution of Motivation and Incentive Systems Research: A Literature Review, http://ssrn.com/abstract=1965646. [13] Jon Christianson, Sheila Leatherman, Kim Sutherland, Financial incentives,healthcare providers and quality improvements:a review of the evidence, QQUIP and the Quality Enhancing Interventions project, ISBN 978-1-906461-02-7. 638 P a g e

A STUDY ON MULTILATERAL DRILLING AND ITS IMPLEMENTATION IN INDIA Nripanka Kalita, M.A. Chowdhury and Minati Das Department Of Petroleum Technology, Dibrugarh University, Dibrugarh, Assam, India. Abstract - Multilateral drilling is a new technology for exploration of hydrocarbons. This technology involves multiple drilling or multi-branched drilling into the formation from a single well to maximize the productivity, and make hydrocarbon recovery more efficient and more profitable. In this review paper based on earlier case studies on the implementation of this technology around the world, different types of multilateral drilling process and configuration have been discussed, to study its applicability in the present hydrocarbon exploration scenario. From the different case studies it is observed that this is the most feasible technique for high productivity in the present day scenario. The study shows an encouraging outlook for future implementation of these techniques in some of oil fields of India has been analyzed and found to be applicable. I. Introduction A multilateral well is a single well with one or more wellbore branches radiating from the main borehole. It may be an exploration well, an infill development well or a re-entry into an existing well. It may be as simple as a vertical wellbore with one sidetrack or as complex as a horizontal, extended-reach well with multiple lateral and sub-lateral branches.[4][6].multilateral configurations include multi-branched wells, forked wells, wells with several laterals branching from one horizontal main wellbore, wells with several laterals branching from one vertical main wellbore, wells with stacked laterals, and wells with dual-opposing laterals. These wells generally represent two basic types: [4] vertically staggered laterals and horizontally spread laterals in fan, Spine-and-rib or dual-opposing T shapes. Like any other well completion, multilateral liners often include external casing packers to ensure zonal isolation or mechanical screens for sand control. Production from individual laterals can be commingled or flow to surface through separate tubing strings. Multilateral well configuration can be: Multi-branched Forked Laterals in to horizontal hole Laterals in to vertical hole Stacked lateral Dual opposing lateral 639 P a g e

Shallow, or heavy oil reservoir Layered reservoir Main well bore Junction Vertically stack lateral Horizontal fanned lateral Dual opposed lateral Low K & naturally fractured reservoir Fig1 - Multi lateral well Fig 1 shows different multilateral wells for different types of reservoirs. For shallow or heavy oil reservoirs, horizontally fanned lateral drilling is suitable, which gives more drainage area and improves the flood pattern or injectivity pattern. In case of heavy oil reserves, keeping two well for injection and one for production can improve the sweep efficiency, hence leading to greater recovery. In layered reservoirs, vertically stacked lateral drilling is suitable for production from different layers of reservoirs. It gives the ease to produce from different layers through a single well head. In low permeability and naturally fractured reservoirs, dual opposed lateral drilling is suitable to tap as many vertical fractures and to increase the drainage area. Multibranched Forke d Lateral into horizontal hole Lateral into vertical hole Stacked lateral Dual opposing lateral Fig2 - Multilateral well configuration [4] Fig 2 shows different configuration of multilateral wells. In this review paper, study has been made an attempt to see the possibilities of implementation of multilateral drilling in some of the oil fields of India. 640 P a g e

PRODUCTIVITY AND RESERVOIR ASPECTS Multilateral wells are drilled to improve the productivity of a field by exploring different zones and layers of reservoirs. The main objective of multilateral well is to provide maximum reservoir contact with the well bore. Multilateral wells are particularly suited for connecting vertical and horizontal features, such as natural fractures, laminated formations and layered reservoirs. Multiple high-angle or horizontal drain holes intersect more natural fractures and often enhance production better than single bore horizontal wells. It reduces frictional pressure losses during production by spreading inflow across two or more shorter lateral branches, hence uniform and productive recovery from the reservoir. Maximizing reservoir contact area increases the wellbore drainage area and reduces pressure drawdown, which mitigates the sand influx and water or gas coning more effectively than that of the conventional vertical and horizontal wells.[3] [4] Multilateral wells are particularly suited for fields with heavy-oil reserves, low permeability or natural fractures, laminated formations or layered reservoirs, hydrocarbons in distinct structural or stratigraphic compartments and matured or depleted reservoir. In matured fields, multilateral wells improve infill drilling by targeting areas that are non economic to produce with a dedicated wellbore. This improves the flow rate from a well and increases recoverable reserves, hence allowing matured reservoirs to be produced economically. Wells with multiple branches help in modifying reservoir drainage in IOR process (water or steam injection projects) by effective flood pattern to improve sweep efficiency; which leads to improvement in productivity. In multilateral wells, when considering the reservoir aspects of improving production, it is important to pay special attention to the skin factor. Lateral branching leads to formation damage and increases skin factor. This formation damage is due to cementing operation, drilling fluid and drilling operation. II. Single HORIZONTAL Vs MULTILATERAL : Fig 3 - Horizontal well CMG Fig 4 - Multilateral well (Stacked lateral) CMG Figure 1 shows the single horizontal well modeled using CMG simulation software and Figure 2 shows the Multilateral Well with two horizontal laterals (Stacked lateral). The main aim of this simulation is to compare the production from both these wells. The model when run showed a difference of 2.00e+9 SCF between their respective cumulative productions over a period of five years. [5] 641 P a g e

Fig 5- Comparison of Cumulative Gas produced from Multilateral & Horizontal Wells [5] Based on the performance curve, we found that the recovery efficiency of multilateral well is more than that of single horizontal well. From the very beginning of the cumulative gas vs. time graph production rate is higher in multilateral well than horizontal well. Though the initial cost of drilling a multilateral well is high, the outcome profit will be more than that of horizontal drilling. Considering all the aspects, multilateral wells will be the most feasible way of hydrocarbon exploration productively and economically. [5] III. REVIEW of successful multilateral wells : Bashkiria field: As with many advances in petroleum technology, the first multilateral well was accomplished by a Soviet drilling engineer Alexander Mikhailovich Grigoryan. In 1953, a unique oil well called simply 66/45 was drilled with turbo-drills in the Bashkiria field near Bashkortostan, Russia. This well ultimately had nine lateral branches from a main borehole that increased exposure to the pay zone by 5.5 times and production by 17-fold, yet the cost was only 1.5 times that of a conventional well. It was the world s first truly multilateral well.[3]. The main bore was drilled to a total depth of 575 meters [1886 ft], just above the pay zone. From that point nine branches were drilled from the open borehole without whipstocks; the window configuration enabled insertion of tools on drill pipe into the sidetracks without instrumentation, with a maximum horizontal reach from kickoff point of 136 meters [447 ft] and a total drainage of 322 meters [1056 ft]. Compared with other wells in the same field, 66/45 penetrated 5.5 times the pay thickness. Its drilling cost was 1.5 times more expensive, but it produced 17 times more oil at 755 B/D [120 m3/d] versus the typical 44 B/D [7 m3/d].[3]. This difference in production rate is due to the more reservoir drainage area in multilateral well. 642 P a g e

66/45 well profile 225 575 Fig6 - An early multilateral well. Drilled in Bashkiria, now Bashkortostan, one of Russia s most prolific regions, the first multilateral well had nine lateral branches that tapped the Ishimbainefti field reservoir.[3] TROLL OLJE field: Troll Olje is part of the Troll gas field,located in the Norwegian sector of the North Sea in 315-340 m of water. Norsk Hydro is developing a 22-26-m oil zone of the Troll Olje oil province and a 13-m oil zone of the Troll Olje gas province to recover 1.33 billion barrels of oil. As of May 2003, 89 wells had been drilled and completed, of which 23 were multilateral wells, including two tri-laterals. The multilateral well concept was introduced on Troll Olje to increase the total drainage area from the existing sub-sea template structures. Further development of Troll Olje includes another 10-15 multilateral wells. These multilateral wells contribute another estimated 90 MMbbl of reserves. Drilling these targets with conventional wells instead would require another six wellhead templates. The tri-lateral X-13 well produces 22,000 b/d from the longest total production screen section in the Troll field, 6,580 m. The draw-down of 0.5 bars is lower than that in traditional horizontal wells, delaying gas break through. This well alone will add 1.5 MMbbl of oil to Troll Olje s total output.[7] PETROZUATA field: Petrozuata, CA, a joint venture company of Conoco Orinoco Inc and Petroleos de Venezuela SA (PDVSA), is developing extra heavy crude in a 74,000-acre block in the Zuata field in the western portion of the Faja del Orinoco in Venezuela. It is one of the world s largest multilateral developments. Each well is designed and geosteered to assure efficient reservoir access based on 3D seismic, vertical stratigraphic well control and detailed geologic facies maps. Petrozuata drilled multilateral wells to develop portions of the reservoir that could be reached from a single surface location to increase well rates while maintaining manageable lateral length. Using a combination of well 643 P a g e

types, designs can be tailored to access the oil in several adjacent areas with fewer pads and wells than the original design would have required. The relative cost of drilling and completing a nine-rib fishbone well was 1.18 times the cost of a single lateral well. The relative costs for other wells were 1.58 times for stacked dual lateral, 1.67 times for dual lateral, and 2.54 times for the more complex crow s foot triple lateral. Petrozuata has found that the increased production rates justify the investment. Multilateral well designs have increased well productivity and estimated ultimate recovery per well, while decreasing cost per barrel of oil developed in the Petrozuata operation.[9] URUCU FIELD: The Urucu field is a remote location situated 650 km southwest of Manaus in the heart of the Brazilian Amazon. The crude produced in Urucu has an average oil gravity of 45 API (Sthener et al. 2010). Daily production is 51,000 BBL and 10 million m3 of natural gas. LPG plants produce 1,500 tons per day. The multilateral-technology wells project ended with the installation of three 9-5/8-in. x 7-in. multilateral TAML Level-4 junctions between 2008 and 2012. All of them were successfully drilled and completed at an average depth of 2,140 m (7,021 ft) TVD, with an average of 1,100 m (3,600 ft) upper lateral horizontals and 1,300 m (4,200 ft) main-bore lateral horizontal drilled and completed. The main and lateral wellbores, including the junction, were cased and cemented. A JST was installed to provide additional support to this cemented junction and also to allow selective access to the main-bore or lateral horizontals at any time during the life of the wells. The typical horizontal well in the Urucu field produces an average of 200 m3 (1260 B/D) drilled in 153 days (on average), with drilling and completion costs of USD 19 million. However, multilateral wells (Dual-lateral wells) had an average production rate of 350 m3 (2200 B/D) of oil with an average drilling time of 202 days per dual-lateral well. The average cost was USD 27 million per dual-lateral well. So by spending only 1.42 times the average drilling cost of a horizontal well, these duallateral wells had an average production rate of 1.75 times more than a typical horizontal well. In addition, the use of multilateral technology in these wells efficiently reduced the environmental impact in a very sensitive area like the Brazilian Amazon by leaving less of a well site footprint, while allowing the drainage of a much larger volume of the reservoir from a single surface location, saving an extra USD 1.5 million (rig location cost), instead of spending USD 3 million (the cost of two rig locations) [7]. IV. FUTURE IMPLEMENTATION IN INDIA: Based on above studies, upgrading conventional vertical well to horizontal well or multilateral well not only saves time and money but also improves productivity by increasing reservoir drainage area, sweep efficiency (heavy oil reserves), improving injectivity pattern, connecting more vertical fractures etc. As shown in fig 1, multilateral drilling techniques can be implemented on some of these following Indian oil fields based on crude oil property (heavy or light), depth of reservoir (Shallow or deep), type of reservoir (layered, naturally fractured or low permeability) etc. Already multilateral drilling technology has been implemented on Mumbai high offshore oil field successfully.[12]. Based on the above study few Indian reservoir descriptions are given below for suitability of multilateral drilling. HEAVY OIL AND SHALLOW RESERVOIRS: 644 P a g e

At Baghawala sandy desert area in Rajasthan's Barmer district 130 million tonnes of 'heavy oil deposits' has been found. It has been planned to dig 60 wells to locate the exact site of heavy oil. The Mangala Area sits in Barmer Basin and is thought to contain oil in place volumes of 3.6 billion barrels (570,000,000 m3), of which 1 billion barrels (160,000,000 m3) are thought to be recoverable [12]. The oil has an API gravity of between 25 and 30, which makes it slightly heavier. However, more importantly, the oil is very waxy. Since this basin will produce heavy and waxy oil so drilling multilateral wells will be suitable and profitable rather than drilling 60 individual wells. The crude is waxy so it is a candidate for thermal recovery or steam injection process. Multilateral well improves the flood pattern or injectivity pattern. Due to improved pattern, injection of steam reduces the waxy nature of the crude leading to increase in recovery. For heavy-oil or other low-mobility reservoirs, lateral drain-holes offer advantages similar to hydraulic fracturing treatments in low-permeability zones. Multilateral well in this basin will provide greater sweep efficiency and greater recovery in case of steam injection or any other EOR process. [11] [12] In addition to improving steam injection, horizontally spread laterals maximize production and improve recovery from heavy-oil deposits and thin, shallow or depleted reservoirs by increasing wellbore drainage area. Ravva oil and gas field of the Krishna-Godavari basin and Rajasthan Block, Barmer, is shallow depth oil and gas reserves, where multilateral (Horizontally spread lateral) wells can be employed for better productivity as described in fig 1.[11][12] Fig7- Horizontally spread multilateral well design LOW PERMEABILITY RESEVOIRS: Low-permeability and naturally fractured reservoirs are frequently associated with limited productivity, so formation anisotropy is a factor in designing multilateral wells. Some reservoirs in Kalol field, Cambay basin are producing heavy oil under low permeability zone. Multilateral well can be a key to recover maximum amount of oil from low permeability zone by increasing the drainage area and tapping trapped oil in naturally fractured zones.[12] 645 P a g e

FIG8- Dual opposed lateral for low permeability reservoir LAYERED RESERVOIR: Mumbai High is a giant offshore oil field discovered in 1974 and located about 160 km westnorthwest of Mumbai city in India. This field is divided into two blocks: Mumbai High North (MHN) and Mumbai High South (MHS). Oil and gas have been discovered in a number of reservoirs, viz., L-I, L-II, L-III, L-IV, L-V, basal classics and fractured basement. Out of these, LII and LIII are the two main limestone oil reservoirs of Miocene age. L- III reservoir is a multilayered limestone reservoir with a gas cap and partial water drive and holds about 94% of the total initial oil in place of the field. Mumbai High is one of the most complex fields in terms of reservoir heterogeneity. The major reservoir (L-III) is multi-layered with 10 limestone sub-layers designated as A1, A2-I to A2-VII, B & C of 2-5 m thick each, separated by 1-5 m shale layers. The Mumbai High South (MHS) Field was put on production in October 1980 and, by September 2010, has produced 263.8 MMt of oil, which is about 26.6% of the in-place oil.[8] [10] Fig9- Vertically stacked lateral 646 P a g e

In layered reservoirs or heterogeneous formations like MHS (L-III pay zone), wells with vertically stacked laterals improve productivity and reserve recovery by connecting multiple pay intervals separated by vertical barriers or permeability contrasts. Simultaneously producing multiple zones helps keep production rates above the economic limit of surface facilities or offshore platforms and prolongs the economic life of wells and fields. V. DISCUSSION AND CONCLUSION As mentioned above as the technology advances, the main motive of petroleum industry becomes reduction of the cost of individual drilling of each reservoir. Upgrading conventional vertical well to horizontal well or multilateral well not only saves the monetary resource but also improves productivity by increasing reservoir drainage area, sweep efficiency, improving injectivity pattern (steam injection in heavy oil reservoirs), connecting more vertical fractures etc. The main aim of using multilateral drilling techniques is to maximize the productivity and recoverable reserves. Moreover multilateral well reduces frictional pressure loss during production, water/gas coning and cost of drilling individual well; minimizes footprint of surface location. However the major disadvantages are high cost, complex drilling operation, difficult well control, bore hole instability and risk of greater formation damage. The key factor in multilateral well is that it improves the flood pattern or injectivity pattern. Due to this injection of steam reduces the waxy nature of the heavy crude leading to increase in recovery. Moreover in lateral drain-holes it offers advantages similar to hydraulic fracturing treatments in low-permeability zones. Thus it provides greater sweep efficiency and greater recovery in case of steam injection or any other EOR process. In upcoming days, multilateral drilling will become the key to explore maximum hydrocarbon with greater profits. Though the initial drilling cost and drilling complications are higher than that of horizontal and vertical drilling, the outcome is way more in multilateral wells. From the earlier reviews we can infer that the drilling cost is 1.5-3 times more than that of horizontal or vertical wells and the production rate is 2-3 times more in case of multilateral wells. Multilateral well already has been implemented on Mumbai high offshore oilfield successfully.[12]. So from this we can infer more multilateral wells in both onshore and offshore oilfields of India which has been found to be a candidate for multilateral drilling technology. Considering both advantages and disadvantages of multilateral drilling techniques, it can be concluded that multilateral well will be the key to explore reservoirs more profitably then vertical and horizontal wells in the upcoming days. REFERENCES 1. New aspects of multilateral well construction by mike Jardon, Mirush kaja, Ramiro paez. 2. Horizontal and multilateral well technology by Sada joshi. Joshi Technologies International, Inc., Tulsa, Oklahoma, USA 3. Bosworth S, El-Sayed HS, Ismail G, Ohmer H, Stracke M, West C and Retnanto A: Key Issues in Multilateral Technology, Oilfield Review 10, no. 4 (Winter 1998) 4. Economic comparison of multilateral drilling over horizontal drilling for Marcellus shale field development by Taha murtuza husain, Leong chew yeong, Aditya saxena. 647 P a g e

5. Sugiyama H, Tochikawa T, Peden JM and Nicoll G: The Optimal Application of Multi-Lateral/Multi- Branch Completions, paper SPE 38033, presented at the SPE Asia Pacific Oil and Gas Conference, Kuala Lumpur, Malaysia, April 14-16, 1997. 6. Hogg C: Comparison of Multilateral Completion Scenarios and Their Application, paper SPE 38493, presented at the SPE Offshore Europe Conference, Aberdeen, Scotland, September 9-10, 1997. 7. Boon or bane: 10 years of multilateral completions by Jim Oberkircher, Ray Smith, Halliburton Energy Services; Ian Thackwray, Woodside Energy Ltd, SPE Annual Technical Conference and Exhibition Denver, Colorado, Oct. 5-8, 2003. 8. Exploitation of Untapped Oil Through Conventional Inclined Wells in a Multilayered Giant Offshore Carbonate Field - A Case Study Uma Goyal, Gulfaraj Anjum, S.K. Anand, S. Chandrasekaran, Search and Discovery Article #20105 (2011). 9. Successful Use of Multilateral Technology to Improve Oil Recovery in the Brazilian Amazon Sandro Mendes, Marcelo Albuquerque, Petrobras, Mario Vento, SPE, Nazildo Batista, Halliburton. SPE Latin American and Caribbean Petroleum Engineering Conference Maracaibo, Venezuela, 21 23 May 2014. 10. www. wikipedia.com 11. www.petrowiki.com 12. www. dghindia.org 648 P a g e

SEMI AUTOMATIC STEREO TRANSFORMATION OF HAND-DRAWN ANIMATION Amit Chaugule, Vipul Barnwal, Chetan Baviskar, Prof. Savita Lade Department of Computer Engineering Rajiv Gandhi Institute of Technology Andheri (W), Mumbai. India Abstract looking ahead, the audience appetite for 3D is slated to grow as 3D content finds new ways to entertain the consumer. 3D production is a young art, and more and more professionals are learning how to produce good 3D. With the increase of films released in 3D, 2D to 3D transformation has become more common. People are likely to go for a 3D movie rather than a traditional 2D movie. The majority of non- CGI stereo 3D blockbusters are converted fully or at least partially from 2D footage. The reasons for shooting in 2D instead of stereo are financial, technical and sometimes artistic. The current 3D transformation technologies used in 3D TVs are not up to the mark and do not provide crisp 3D content like the one we get using menial work 3D transformation process. We propose a method to convert 2D hand drawn frame sequence into stereoscopic 3D, without labor-intensive or manual depth assignment or 3D geometry reconstruction. By utilizing the T-junction cue available in hand drawn animation. We initially order the different regions in each frame, these regions may not necessarily be of exact depth, but may provide convincing 3D visuals. Further we try to introduce depth to the scene. Keywords ordering, stereo conversion, T-junction, view angle based render. 1. Introduction 2D to 3D movie conversion is the process which is also called as converting 2D flat picture into stereo 3D which in almost all cases is stereo, so it is the process of creating two images from one 2D image one for each eye. 3D images are obtained by looking at one image and one image from right eye. Normally 3D movies are created using 3D camera containing two lenses. Some of this can be determined naturally. For example: brighter objects are closer; larger objects are closer. Many 3D TVs, Blu-ray players and outboard frame sequence processors (like the recently-reviewed 3D-Bee) can do 2D-to-3D transformations in real time, but the results are erratic at best. While these devices will no doubt get better with time, a proper transformation job requires human interaction and guidance. We propose a method to convert 2D hand drawn frame sequence into stereoscopic 3D, without menial depth assignment or 3D geometry modelling. Using T-junction information available in hand drawn frame sequence we initially arrangement of the different objects in the frame sequence, these objects may or may not be in the same arrangemant, but may provide compelling stereo 3D. Further we try to introduce deepness 649 P a g e

to the scene. Deepness to the scene in our case is synthesized using the motion or the displacement of the region. Other deepness calculation method may also be implemented as per the complexity of the scene. 2. Existing System 2.1 Provide Z-map Most semiautomatic methods of stereo transformation use Z-map and depth-image-based rendering. The idea is that a separate auxiliary picture known as the "Z-map" is created for each frame which suggest position of object on the z-axis in scene. The Z-map is a grayscale image having the same resolution as the actual 2D image. Various shades of gray suggest the deepness of every part of the frame. z-map can produce proper 3D illusion but it may not work properly in case semi transparent objects nor does it allow explicit use of blockage, we need to apply different methods to overcome this problems. The major steps of depth-based transformation methods are: Deepness allocation How much depth should be there in the scene so that viewer must be comfortable by watching video. Which object should lie on which plane. How much level of detailing should be there in the scene. This detail depend upon cost of conversion. Z-map creation- Each separated object must be given position on the z- axis which means each object must be assign a gray scale value on z map. This is the process of visualizing every arrangement on every change in ordering. Particular scene which contains high level of noticeable details must be converted using geometric reconstruction. Normally the original image is kept as image for one eye and the image for other eye is constructed this decision is based on the conversion cost because generation of two images from the original image for both the eye is costlier than generating only one image. 2.2 3D Geometry Reconstruction In this method initially a 3D model resembling the object in the scene is constructed then the object is project over a 3D model and capture two images from slightly different angle in order to get two images for two eyes. This is possible for only in the case of static objects like steady mountain or person sitting on the bench but not applicable for moving objects like man is walking or fluids like gases. 2.3 Automatic System Using different types of motion it is possible calculate the position of object on Z-axis.For example, when you are travelling through a train when you look out of window the tress appear to be moving fast and mountains stays at its place. what this suggest us faster moving objects are closer and slow moving objects are farther. Another example is, if you see railway track they appear to be merge at certain point from that we get to know about deepness. and another example is, if you capture focused image from that we will get to know about foreground and background. 3. Proposed System 3.1 Edge Detection We use a simple technique where we apply edge detectors to obtain edges. First we multiply the luminance component which a multiplicative factor mf which changes the difference in neighboring pixel. mf is provided 650 P a g e

by the user, then we apply sobel operator with a thresholding factor tf. The detailing of the segments or the 3D depth depends on the thresholding factor tf and multiplicative factor mf. Fig. 3.1 edge map 3.2 Region Detection Initially take edges input where edge is represented as logical 1 and background represented as logical 0. Invert the image (or initially take the above values inverted).label each segment. Use area opening operation in order to remove small areas. Then take each labeled segment and then perform erosion followed by dilation. Erosion removes the edges in the complete image. And the only selected labeled image is remaining. Then perform dilation to recover lost edge of the labeled image. Perform this operation in order to include edges into segments. And also remove unwanted small areas. 3.3 Junction Detection (T-JUNCTION) Wherever three regions meet there forms a T-junction. Find image gradient (to travel only through the edges). Travel through every pixel, W(x, y) in the gradient. Check for adjacent pixel W(x+1,y), W(x,y+1), W(x+1,y+1). Check if we get three different intensities. If w(x, y) satisfies above criteria register it. This w(x, y) defines the T-junctions Fig. 3.2 labeled edge map 651 P a g e

Fig. 3.3 region generation Fig. 3.4 Z-map 3.4 Junction Analysis Traverse through the registered T junctions. Clip a window of size N x N. Count number of pixel of each intensity in the clipped window and arrange the intensities in descending order- sorted form according to respective count. The first intensity represents foreground and remaining two represents background. N x N window might contain small unwanted regions too. But we ignore those regions assuming that there is enough other T-junction who would participate in our voting system. Based on the voting system if in case there are vague t-junction, discard those T-junctions. And based on our assumption other t-junction correct those t- junction ordering. 3.5 Local Ordering Apply depth first search. If a loop is formed remove the loop. Create a empty register r and solution and a variable level=0. Add all the nodes that have only the nodes in register r as the root in register solution. Number all the nodes in solution. level = level. Increment level. Add all the new entries in solution to register r. Repeat steps till all the nodes are added. 3.6 Global Ordering There may be inconsistency over the regions throughout the frames. Here we manipulate local orders of continuous frames to get a sense of consistency over the regions i.e. to keep an immobile or comparatively immobile region to be on a specific depth and similarly mobile frames to move imp portions to their speed and motion. In the initial stage of region detection once we find the region maps. We change the region maps in the following image based on the previous image. We use the max [intensity] in original image based on labeled image. There are cases when conflict occurs in max [intensity]. for this problem we maintain a register. where we store conflicts and assign a unique number to it. Further we check regions.centroid and area and keep track of it. Then change the solution vector where we actually store the levels of the regions. A level is the position of the region in the z-axis. Max level defines regions in positive z-axis and min level defines region in negative z- axis 3.7 Final Render First take the map which contains the depth knowledge. Stretch the original image twice its size stretch the map twice its size. Using the equation 3.1 generate the points in the new image based on original age. We have developed a simple algorithm to generate new image i.e. the left and right image. Our algorithm uses simple stretch and fill technique. The input to the algorithm will be the viewing angle. To generate red cyan image.take 652 P a g e

red channel from left image. take green channel and blue channel from right image concatenate outputs of step (i) and (ii). To generate side by side images. Stitch left image and right image Fig. 3.5 final render 4. Equation We use the following algorithm to render the left and right image. The equation requires viewing angle and the Z-map. The Z-map is generated using t-junction and depth through motion.here we intent to obtain the rendeered image from the original image using view angle based render algorithm.the user initially provides theta which in our case is 30 degree. Here y represents y th column ie. The y th co-ordinate of the pixel. Map(x,y) represents the position in the z axis.here z axis is an imaginary axis. We use the following trignometric knowledge to compute the distance k. which is the kth column or the kth co-ordinate in the rendered image.in short this algorithm predictes the position of pixel in rendered image. Fig. 4.1 view angle based render Given h=y* 653 P a g e

5. Results Fig. 5.1 sample input sequence Fig. 5.2 output sequence for sequence in fig 4a) without consistency and global ordering Fig. 5.5 sample image Fig. 5.6 after converting Fig 5.5 6. Conclusion 3D transformation has evolved over past many years and it has a great scope. With the increase of films released in 3D, 2D to 3D transformation has become more common. With such a demand for the 3D tv s industry, it is necessary to have 2D movies made into 3D.Our first contribution is ordering technique and second contribution is rendering the left and right image by eliminating inpainting stage also eliminate use of 3-D geometric reconstruction. No use of 3D geometry reconstruction will reduce the complex hardware requirement. 654 P a g e

Reduction in menial work manual depth assignment will reduce the cost of 3D conversion. Using our approach of converting hand drawn frame sequence into stereoscopic 3D, will not only help shape cartoon industry but also help grow 3Dtv industry. This approach will somehow act as base to convert true color images into stereoscopic 3d in near future. REFERENCES [1] Jia. Z., Gallagher, A., Chang, Y., And Chen, A learning-based framework for depth ordering, In Computer Vision and Pattern Recognition (CVPR), 2012 IEEE Conference on, IEEEA, 294?301. T. 2012. [2] Apostoloff, N., And Fitzgibbon, A. 2005. Learning spatiotemporal T-junctions for occlusion detection, In Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer Society Conference on, vol. 2, IEEE, 553.559.2005. [3] Dimiccoli, M., And Salembier, Exploiting t junctions for depth segregation in single images. In Acoustics, Speech and Signal Processing, 2009. ICASSP 2009. IEEE International Conference on, IEEE, 1229?1232.2009. [4] Dimiccoli, M., And Salembier, Hierarchical region based representation for segmentation and ltering with depth in single images, In Image Processing (ICIP), 2009 16th IEEE International Conference on, IEEE, 3533?3536. 2009. [5] Xueting Liu, Xiangyu Mao, Xuan Yang, Stereoscopzizing cel animation, ACM transcation on Graphics(Siggraph Asia 2013 issue), 223:1-223:10. 2013 655 P a g e

ARTIFICIAL EYE (BIONIC EYE) Shaily Tomar and, Richa Sharma Dept. of Electronics and Communication JIET Group of Institutions Abstract--- India is now home to the world's largest number of blind people. In 37 million people across the globe 15 million blind people are from India. 75% of these are cases of avoidable blindness. On the other hand, while India needs 2.5 lakh donated eyes every year, the country's 109 eye banks in which 5 located in Delhi manage to collect a maximum of just 25,000 eyes, 30% of which can't be used. Meanwhile, shortage of donated eyes is becoming a huge problem. In 15 million blind people in India, three million that is 26% of whom are children who suffer due to corneal disorders. But only 10,000 corneal transplants are being done every year because of shortage of donated eye. The bionic eye aims to restore basic visual cues to people suffering from eye diseases such as retinitis pigmentosa, which is a genetic eye condition.a video camera fitted to a pair of glasses will capture and process images. These images are sent wirelessly to a bionic implant at the back of the eye which stimulates dormant optic nerves to generate points of light (phosphenes) that form the basis of images in the brain. Thus even blind people can have vision. Keywords--- Bionic Eye, Artificial Silicon Retina, Implantation I. INTRODUCTION BIOTECHNOLOGY has become the fastest-growing area of scientific research, with inventions of new devices. Technology has done wonders which help the mankind. Prosthetics use to help overcome handicaps. A training system called brain port is letting people with visual and balance disorders bypass their damaged sensory organs and instead send information to their brain through the tongue. Now, a company called second sight has received FDA approval to begin U.S. trials of retinal implant system that gives blind people a degree of vision. II. THE HUMAN EYE We are able to see because light from an object can move through space and reach our eyes. Once light reaches our eyes, signals are sent to our brain, and our brain deciphers the information in order to detect the appearance, location and movement of the objects we are sighting at. 656 P a g e

The whole process, as complex as it is, would not be possible if it were not for the presence of light. Without light, there is no world. The human eye is the organ which gives us the sense of sight, it allows us to learn about the surrounding world than any of the other senses. Fig. 1: How Human Eye Works The eyeball is present in a protective cone-shaped cavity in the skull called the orbit or socket and measures approximately one inch in diameter. The orbit is covered by layers of soft, fatty tissue which protect the eye and enable it to turn easily. The important part of an eye is retina. The retina lies at the back of the eye and it acts as though the film in a camera act by receiving and processing everything. III. THE BIONIC EYE Bionic Eye is an artificial eye which provokes visual sensations in the brain by directly stimulating different parts of the optic nerve. Bionic eye consist of electronic systems which consist of image sensors, processors, receivers, radio transmitters and retinal chips. There are also other experimental implants that can stimulate the ganglia cells on the retina or the visual cortex of the brain itself. Technology paved way through a bionic eye to allow blind people to see again. Fig. 2: Bionic Eye It comprises a computer chip which is kept in the back of the individual's eye, linked up using a mini video camera built into glasses that they wear. Images captured by the camera are Beamed to the chip, which translates them into impulses that the brain can interpret. 657 P a g e

Although the images produced by the artificial eye were far from perfect, they could be clear enough to allow someone who is otherwise blind to recognize faces. The breakthrough is likely to benefit patients with the most common cause of blindness, macular degeneration, which affects 500,000 people. This occurs when there is damage to the macula, which is in the central part of the retina where light is focused and changed into nerve signals in the middle of the brain. The implant bypasses the diseased cells in the retina and stimulates the remaining viable cells. IV.WORKING OF THE BIONIC IMPLANT A bionic eye implant that could help restore the sight of millions of blind people could be available to patients within two years. Fig. 3: How the Bionic Eye Implant Works This device is 2 millimeters across and contains some 3,500 micro photodiodes which is placed behind the retina, this collection of miniature solar cells is designed to convert normal light to electrical signals, which are then transmitted to the brain by the remaining healthy parts of the retina. A Belgian device has a coil that covers around the optic nerve, with only four points of electrical contact. By shifting the phase and varying the strength of the signals, the coil can stimulate different parts of the optic nerve, rather like the way the electron guns in TVs are aimed at different parts of the screen. The video signal senders from an external camera and are transmitted to the implant through a radio antenna and microchip under the skin just behind the ear. Implants of a microchip, smaller than the head of a pin and about half the thickness of a sheet of paper were used to remove blindness. The eye-position monitor controls the image camera's orientation. If the image-acquisition camera is not mounted on the head, compensation for head movement will be needed. Finally, if a retinal prosthesis is to receive power and signal input from outside the eye via an IR beam entering the pupil, the transmitter must be aligned with the intraocular chip. The beam has played two roles: one is to sends power, and another is to send pulse - or amplitude - modulated to transmit image data. Using the control of eye movement, the main imaging camera for each eye can swivel in any direction. Each of these cameras--located just outside the users' field of view to avoid Blocking whatever peripheral vision they might have captures the image of the outside world and transmits the information through an optical fiber to a signal-processing computer worn on the body. 658 P a g e

The Argus II system uses a spectacle-mounted camera which is used to send information to electrodes in the eye. Patients who tested less-advanced versions of the retinal implant were able to see light, shapes and movement. The function of Bionic eye is to take real-time images from a camera and convert into tiny electrical pulses that help the blind eyes to see. 1: Camera which is implanted on glasses helps to view the image. 2: Signals are sent to hand-held device 3: The information which processed is sent back to glasses and wirelessly transmitted to receiver under surface of eye. 4: Receiver sends information to electrodes in retinal implant. 5: Electrodes stimulate retina to send information to brain Fig. 4: Retinal Implant Retinal implants can partially restore the vision of people with particular blindness caused by diseases such as macular degeneration or retinitis pigmentosa. About one and half million people worldwide have retinitis pigmentosa, and one in ten people over the age of fifty five have age-related macular degeneration. Both diseases cause the retinal cells which process light at the back of the eye to gradually diminish. The new device invented work by implanting an array of tiny electrodes into the back of the retina. A camera is used to capture pictures which consist of a processing unit about the size of a small handheld computer and worn on a belt helps to convert the visual information into electrical signals. These are then sent back to the glasses and wirelessly on to a receiver just under the surface of the front of the eye, which in turn feeds them to the electrodes at the rear. Growing Dots First-generation, low-resolution devices have already been fitted to six patients. Brain Change The new implant has a higher resolution than the earlier devices, with 60 electrodes. 659 P a g e

V. RETINAL PROSTHESIS SYSTEM Fig. 5: Second Sight Second Sight Medical has just received USFDA Investigational Device Exemption (IDE) to begin clinical trials for their Argus II Retinal Prosthesis System. At Second Sight, their retinal prosthesis uses an array of electrodes to stimulate the retina. It restores a low level of vision in patients with degenerative diseases. Their first implant had sixteen electrodes; the new Argus II has 60 electrodes. The Argus II implant consists of an array of electrodes that are attached to the retina and used with an external camera and video processing system to provide a rudimentary form of sight to implanted subjects. An IDE trial of the first generation implant (Argus 16), which has sixteen electrodes, is ongoing at the Doheny Eye Institute at the University of Southern California. The Argus 16 was implanted in six Patients between 2002 and 2004 and has enabled them to detect when lights are on or off, recognize an object s motion, count items, as well as locate and differentiate basic objects in the surrounding. Fig. 6: Argus II The next generation Argus II retinal stimulator is designed with 60 controllable electrodes, which should provide implanted subjects with higher resolution images. Second Sight remains the only manufacturer with an actively powered permanently implantable retinal prosthesis under clinical study in the United States, and the technology represents the highest electrode count for such a device anywhere in the world. 660 P a g e

Fig. 7: Working of Artificial Eye Fig. 9: Artificial Eye VI. CERAMIC PHOTOCELLS Scientists at the Space Vacuum Epitaxy Centre (SVEC)[9] based at the University of Houston, Texas, uses a new material, comprising tiny ceramic photocells that detects incoming light and repair malfunctioning human eyes. Scientists at SVEC are conducting preliminary tests on the biocompatibility of this ceramic detector. The artificial retinas constructed by SVEC consist of 1, 00,000 tiny ceramic detectors, each1/20th the size of a human hair. The assemblage is so small that surgeons can t safely handle it. So, the arrays are attached to a polymer film one millimeter in size. After insertion into an eyeball, the polymer film will simply dissolve leaving only the array behind after a couple of weeks. ADVANTAGES It helps to correct the vision. There is no necessity to suffer from long and short sights. It can be easily implanted It is the one approved by FDA DISADVANTAGES There are 120 million rods and 6 million cones in the retina of every healthy human eye. Creating an artificial replacement for these is a risky task. Si based photo detectors have been tried in earlier attempts. But Si is toxic to the human body and reacts unfavorably with fluids in the eye it cost about 30,000$ It will not be helpful for glaucoma patients. VII. CONCLUSION Restoration of sight for the blind is no more a dream for people today. Bionic Eyes have made this true. Though there are a number of challenges to be faced before this technology reach the common man, the path has been laid. This paper has tried to present the concept of Artificial Vision called Bionic Eyes. It is just a matter of time, may be 4-5 years that the blind will be able to see through these Bionic Eyes, with thanks to Science and Technology. 661 P a g e

REFERENCES 1. http://news.bbc.co.uk/2/hi/health/4411591.stm 2. http://www.docstoc.com/docs/4340408/bioniceye 3. http://seminarprojects.com/thread-bionic-eyes-for-the-blind-full-report 4. http://news.bbc.co.uk/2/hi/health/4411591.stm 5. http://www.fanboy.com/2007/02/page/3 6. http://www.medgadget.com/2007/01/second_sight_me.html 7. http://science.howstuffworks.com/life/human-biology/bionic-eye1.htm 8. http://www.123seminarsonly.com/seminar Reports/002/64285761-Bionic-Eye12.pdf 9. Bionic Eye Technology: An Advanced Version of Artificial Vision, Academia, Volume 2, Issue 7, July, 2012. 10. Effect of shape and coating of a sub retinal prosthesis on its integration with the retina. Butterwick, P. Huie, B.W. Jones, R.E. Marc, M. Marmor, D. Palanker. Experimental Eye Research 11. W.H. Dobelle, ``Artificial Vision for the Blind by Connecting a Television Camera to the Visual Cortex,'' ASAIO Journal (American Society for Artificial Internal Organs), January - February 2000. 12. Amedi, F. Bermpohl, J. Camprodon, S. Fox, L. Merabet, P. Meijer and PascualLeone, ``Neural correlates of visual-to-auditory sensory substitution in proficient blind users,'' poster presentation at CNS 2005 (12th Annual Meeting of the Cognitive Neuroscience Society) in New York, USA, April 11, 2005, and at the 57th Annual Meeting of the American Academy of Neurology (AAN 2005), Miami Beach, Florida, USA, April 10 and 12, 2005 662 P a g e